499 results about "Data Origin" patented technology

The present invention relates to a system and method for cross-domain predictive modeling using Bedrock-based foundation models and blockchain-anchored data. The invention integrates large-scale foundation model reasoning with distributed ledger-based data provenance to enable verifiable, secure, and explainable predictive analytics across heterogeneous domains such as finance, healthcare, logistics, and environmental systems. The system comprises a data ingestion unit for receiving and normalizing multi-domain datasets, a blockchain anchoring unit for generating cryptographic hashes and recording data provenance into a distributed ledger, a cross-domain harmonization processor for aligning heterogeneous feature representations into a unified latent space, a foundation model processor configured to execute Bedrock-based predictive inference with adaptive domain contextualization, a verification processor for validating predictions against blockchain-anchored ground truths, and a governance processor for maintaining immutable audit trails of model evolution.

The invention belongs to the technical field of computing resource scheduling, and particularly relates to an intelligent scheduling method and system for Huaan Atlas heterogeneous computing resources based on dynamic load awareness. The method comprises the steps that the real-time state of multi-dimensional hardware data is collected, and a basic data source is provided for subsequent steps; dynamically adapting tasks and hardware characteristics through a matching degree matrix, modeling aiming at various basic data, and constructing a state vector required by reinforcement learning; predicting a fault risk score through a lightweight prediction model deployed at each computing node; and a deep Q network is adopted as a model architecture, a state vector and a fault risk score are input, reinforcement learning training is performed through a reward function in a multi-target vector form, a final scheduling model is obtained, and a task allocation decision is output. The problems that in the prior art, the hardware state cannot be sensed in real time, hardware characteristic matching is ignored, consequently, the computing resource utilization rate is insufficient, and fault recovery is passive are solved.

The invention discloses an industrial building business service management method and device and a medium, and belongs to the technical field of building service management, and the method specifically comprises the steps: constructing a multi-dimensional data fusion model; defining cost, progress, quality and security dimension data ranges and acquisition nodes, and storing the data ranges and the acquisition nodes in a model database after standardization; based on the standardized data of the multi-dimensional data fusion model, analyzing a dimension internal relationship by using an association rule mining algorithm, defining a dimension mapping relationship, and integrating to form an association rule base; establishing a multi-dimensional linkage analysis module; the module is established by taking a model database as a data source and a multi-dimensional data association rule base as a logic support, and a threshold triggering mechanism and a cross-link data calling function are set; constructing a business collaborative decision support module; business requirements of multiple departments are integrated; a multi-dimensional data visualization monitoring platform is built; and designing a linkage trend chart based on a result of the multi-dimensional linkage analysis module.

The invention belongs to the technical field of knowledge Token generation and traceability, and discloses a knowledge Token generation and traceability method based on structured metadata, which is characterized in that a current knowledge scene is quickly matched and a core field is dynamically adjusted through few-sample learning in combination with various technical models and a preset industry field model library; a mixed semantic segmentation mode is adopted, knowledge fragments are completely segmented according to business logic, then a refined element extraction strategy is matched, three types of key elements including entities, relationships and clauses are accurately extracted, and errors are reduced through multi-dimensional verification; originally dispersed PDF, audio and video and other unstructured knowledge can be converted into standardized assets with anchoring information. Meanwhile, a multi-dimensional verification process is designed, comprehensive verification is carried out from a data source, access permission and logic consistency to a health state, it is ensured that a knowledge source can be checked, authenticity can be distinguished, and the information distortion risk during large model calling is reduced.

The invention provides a time sequence data storage method and system, electronic equipment and a storage medium, and the method comprises the steps: obtaining time sequence data and metadata information of the time sequence data, the metadata information comprising a first point location identifier, a data source and a data type corresponding to the time sequence data; according to the data source and the data type, determining a storage strategy of the time series data, including a target storage period and a target storage position; according to the storage strategy and the first point position identification, a target storage instance corresponding to the time sequence data is matched from a storage instance set, all storage instances in the storage instance set are combined and constructed according to different storage strategies and the point position identification, and all the storage instances are parallel storage instances; and storing the time sequence data to a time sequence database through the target storage instance. By determining a dynamic storage strategy driven by metadata and a parallel storage instance, the problem of performance bottleneck during high-concurrency writing of massive time series data is solved, and the real-time performance and reliability of time series data storage are ensured.

The invention relates to a real-time process monitoring method based on data analysis, and the method comprises the following steps: S1, building a three-dimensional component priority evaluation model based on a component operation scene type, a real-time resource occupancy rate and a data dependence degree, and dynamically matching an acquisition strategy; s2, a cleaning rule is adapted according to a data source, a feature extraction dimension is adjusted in combination with process dynamic features, and an improved time sequence decomposition algorithm is used for separating data trends, fluctuations and abnormal residual errors; s3, constructing an exclusive baseline sub-model by using an online learning algorithm according to a scene label, and establishing a scene switching mechanism; s4, in combination with component interaction anomaly features, an anomaly level is judged through a mixed detection model; s5, on the basis of exception processing and user feedback, constructing an error correction model optimization parameter; and S6, generating a report containing an abnormal propagation path, and triggering hierarchical collaborative response of the associated component. The invention aims to solve the problems of single acquisition dimension, no scene adaptability in preprocessing, incomplete abnormal detection and the like in the existing monitoring technology.

The invention discloses a multi-type energy scheduling method based on a trusted data space, and the method comprises the steps: accessing a multi-type energy provider through a block chain node, and constructing a cross-domain trusted data space; obtaining energy data of each party; processing to obtain an energy data set; calling a federated learning model, training a load prediction neural network, and predicting to obtain a future load demand; constructing a multi-target optimization model by taking the maximum consumption efficiency and the minimum energy abandoning rate of each type of energy as targets, and solving the multi-target optimization model based on the predicted future load demand and the output curve of each type of energy to obtain an optimal scheduling strategy; and storing the optimal scheduling strategy on the chain, and obtaining the optimal scheduling strategy on the chain by each type of energy providers to perform energy scheduling. A decentralized trusted data space is constructed through a block chain technology, data source traceability and non-tampering operation are realized, a centralized platform trust risk is avoided, and data privacy zero leakage during cross-domain sharing is ensured by adopting privacy calculation.

The invention discloses a semantic response method and system based on a large model and a knowledge interface, belongs to the technical field of artificial intelligence, and aims to solve the technical problem of how to realize reliable answering and analysis in a factual question and answer scene. Storing the standardized description and the vectorized description of the function description of each knowledge interface into a Milvus vector database; matching related knowledge interfaces from the Milvus vector database to serve as candidate interfaces based on user intentions for current user input serving as first questions, and taking knowledge interfaces corresponding to associated historical dialogues as candidate interfaces of the current user input for current user input serving as non-first questions; extracting information corresponding to the required parameters from the current user input, filling the extracted information into an interface document of the candidate interface, and verifying the filled candidate interface; structured data is returned based on the knowledge interface, the structured data is generated into a natural language answer through a large language model, and a data source is attached.

The invention belongs to the technical field of large models and data agents, and particularly relates to a question answering method based on a large model and a data agent, and the method comprises the steps: receiving and preprocessing a user question, and generating a question data set; judging whether external data support is needed or not by utilizing a large model, and if so, dynamically generating an agent calling strategy; real-time data acquisition, domain knowledge analysis, multi-source data fusion and logical reasoning verification intelligent agent cooperative execution tasks are scheduled according to a strategy; fusing the traceability information returned by the intelligent agent and the knowledge of the large model to generate a preliminary answer, and outputting the preliminary answer after interpretable optimization; and finally, optimizing the case library according to user feedback and iterating an agent calling strategy. All data agent processing results are attached with full-link traceability information, data sources and reasoning processes are presented in an explanatory optimization mode, and a user can verify answer basis.

The invention relates to the technical field of natural language processing, and discloses a diagnosis and treatment quality evaluation method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring multi-source information data of a target patient in a treatment process; wherein the multi-source information data comprises target dialogue data and associated diagnosis and treatment data; performing reconstruction based on the multi-source information data to obtain an actual diagnosis and treatment thinking chain; wherein the actual diagnosis and treatment thinking chain is used for representing a diagnosis and treatment reasoning process when a doctor inquires a target patient; performing comparative analysis based on the actual diagnosis and treatment thinking chain to generate a diagnosis and treatment quality evaluation result for the target patient; wherein the diagnosis and treatment quality evaluation result is used for performing comprehensive quality control on the reasonability of the medical treatment process. By acquiring the multi-source information data containing the target dialogue data and the associated diagnosis and treatment data, the defect of single source of traditional quality control data is made up, and finally reasonable and comprehensive quality control can be performed on the whole diagnosis and treatment process.

The invention discloses a meteorological equipment maintenance knowledge base construction method and device and a medium, and the method comprises the steps: obtaining multi-modal data related to meteorological equipment maintenance, and synchronizing structured data in the multi-modal data to a knowledge base management platform through a Restful API interface; the multi-modal data further comprises unstructured data; analyzing the unstructured data to extract image-text features, and asynchronously transmitting the crawled webpage content to a knowledge base management platform through a message communication protocol; constructing a vector index and knowledge graph association relationship on a knowledge base management platform, and labeling a multi-modal semantic tag of an equipment type and a fault phenomenon; a natural language query request of a user is received, knowledge base fragments are retrieved through keyword retrieval and vector similarity calculation, retrieval results are input into a large language model to generate equipment maintenance suggestions, and associated data sources are output.

The invention relates to the technical field of power system data processing, and discloses a meteorological data collection method based on a power grid prediction demand. The problems that existing meteorological data sources are high in isomerism, the preprocessing process is repeated, the labor amount is large, unified standards are lacked, and the self-repairing capacity is insufficient are solved. The method comprises the following steps: constructing a multi-source meteorological data access layer and carrying out unified space-time reference conversion; analyzing a power grid prediction task demand, and dynamically matching and splicing the preprocessing sub-architecture units; executing the customized preprocessing assembly line to generate a numerical mapping graph; performing logic consistency repair on missing or abnormal data; and outputting the numerical mapping graph after integrity restoration to a power grid prediction model interface. According to the technical scheme, unification and adaptation of multi-source heterogeneous meteorological data at the acquisition end according to needs can be achieved, data processing flexibility and expandability are improved, data integrity and physical consistency are guaranteed, and power grid prediction precision and response speed are improved.

The invention discloses a business flow site selection optimization method based on multi-source data fusion, and relates to the technical field of big data processing and multi-source data fusion, and the method comprises the following steps: obtaining multi-source data of a to-be-detected region, constructing multi-source features, carrying out the cross-modal fusion of the multi-source features through an adaptive weight distribution mechanism, and generating a comprehensive feature representation; based on comprehensive feature representation, passenger flow trends and causal contributions of the candidate points in different scenes are output through a causal inference method; taking the passenger flow trend and the causal contribution as input, establishing a multi-objective optimization model, performing solving, and outputting a comprehensive score and a sorting result of the candidate points; constructing an explanatory index matrix according to the comprehensive score and the sorting result in combination with a causal contribution result, and outputting a site selection scheme through a visual platform; according to the method, the problems of single data source and incomplete candidate point description in the prior art are solved by fusing the spatial geographic data, the space-time traffic data and the consumption behavior data.

The invention provides a security optimization encryption method, system and device for data in a facility agriculture scene and a medium, and belongs to the technical field of agricultural informatization security, and the method comprises the steps: carrying out the dynamic division of security levels of data collected in a facility agriculture environment according to the data source device type, data content sensitivity and real-time requirements; distributing an encryption algorithm for the collected data according to a mapping relation between the security level and the encryption algorithm; secret keys are generated and managed based on an identity encryption mechanism, and then secret key fragmentation storage is carried out in combination with a trusted execution environment; when encryption is needed, whether the current trusted execution environment is safe or not is judged, and when the current trusted execution environment is safe, secret keys of the fragments are recombined, the multi-source heterogeneous data are encrypted through the distributed encryption algorithm, and encrypted data are generated; and carrying out secure transmission on an adaptive communication protocol after packaging the encrypted data. According to the method, the data security level is dynamically divided, the encryption algorithm is flexibly selected, the resource consumption is reduced, and the real-time performance and security are improved.

The invention discloses a data quality monitoring and intelligent blocking system, system and product based on a rule engine, which are used in the field of data management. The system comprises a basic website, an intermediate transfer center and a headquarters data center. The basic network point obtains the express data, performs data legality verification, blocks illegal data, and transmits legal data to the intermediate transfer center; the intermediate transfer center obtains input data, calibrates a first input link, performs analytical verification, blocks data not passing the verification, sends a blocking feedback signal to a data source, and transmits the data passing the verification to the headquarters data center; and the headquarters data center obtains input data, calibrates a second input link, associates the second input link with the first input link, and provides data reports of different dimensions based on a preset analysis dimension after storage. According to the method and the device, the express data is monitored in a full-process coverage manner, the problem data is prevented from being propagated to the downstream, and the express data is comprehensively, conveniently, efficiently and reliably managed and controlled.

The invention relates to a data consanguinity analysis method, device and equipment, a medium and a product in the technical field of data processing. The method comprises the following steps: extracting a source path and a processing mode of a field according to an original data script in a data blood relationship link, and constructing a field-level data blood relationship map; matching field contribution degree levels based on the field processing mode and outputting a field level contribution score matrix; and based on the field level contribution score matrix and a preset aggregation rule, outputting a contribution degree result of the field, table level comprehensive contribution and a data blood relationship network. According to the embodiment of the invention, factors such as the source path and the processing mode of the field are combined, the table-level comprehensive contribution value is further summarized and generated through the constructed field-level basic weight model, and quantitative evaluation of the key node in the data source path is realized.

The invention provides an intelligent water conservancy monitoring method and system based on the Internet of Things. The method comprises the steps of obtaining heterogeneous water conservancy data of various sources; preprocessing and standardizing the heterogeneous water conservancy data from various sources; constructing a water conservancy domain knowledge graph, and performing multi-dimensional data fusion to generate comprehensive water conservancy situation data; based on the comprehensive water conservancy situation data, constructing a water conservancy digital twinborn body, and realizing real-time driving and hydrological and hydraulic process simulation of the water conservancy digital twinborn body; and multi-modal deep analysis is performed on the comprehensive water conservancy situation data and the water conservancy digital twin, and interpretable intelligent decision support information is generated by using an artificial intelligence technology. According to the method, the problems of scattered data sources, high isomerism, relatively low intelligent degree and insufficient information presentation modes of a water conservancy management system are solved by acquiring and preprocessing the multi-source heterogeneous water conservancy data, fusing and cognizing the water conservancy data, twinning and simulating the water conservancy system and performing analysis and decision intelligent support.

The invention provides a multi-organization-level salary data acquisition method, and relates to the technical field of enterprise informatization management, and the method comprises the steps: a group headquarters defines a statistical scheme containing basic information, statistical items, data sources and verification rules, and issues the statistical scheme to lower-level units step by step; each subordinate unit creates a collection scheme of the level according to the actual situation, and establishes a mapping relation with a superior scheme; the last-level unit defines a salary scheme, a salary item and a statistical item mapping relation, report data is generated, and data verification is carried out according to a verification rule of a statistical scheme. And after receiving the data, the superior unit checks, integrates and verifies the data, and the process is repeated step by step until the data is summarized to the headquarters of the group. And the group headquarters audit and integrate the final data to generate statistical data, thereby realizing all-staff, all-level and whole-process data acquisition and management and control.

The invention discloses a circuit board manufacturing method and electronic equipment, and relates to the technical field of production data collaborative analysis, and the method comprises the steps: firstly building a process influence coupling model based on a single board time sequence data sequence corresponding to the whole flow of a production line in a historical production process; the process influence coupling model can quantify the dynamic influence capability of each process on the subsequent quality state. Therefore, the process influence coupling model and the credibility of the data source equipment of each process can be combined to determine the fusion weight expressing the influence causal relationship of each process and the reliability of the data. Thus, the fusion weight is utilized to fuse the quality features of each process of the current batch, a comprehensive quality representation vector corresponding to the whole-process collaborative state can be obtained, and a defect propagation path diagram matched with the comprehensive quality representation vector is constructed. By means of the defect propagation path diagram, the target working procedure which causes the circuit board to generate defects at present can be determined from the production line, and the target working procedure is optimized, so that the overall process performance can be effectively improved.

The embodiment of the invention provides a vehicle control method which comprises the steps that first road condition data and second road condition data of a preset road are obtained, the first road condition data are from a platform data source, and the second road condition data are from a vehicle-mounted data source; performing risk classification processing based on the first road condition data to obtain a target risk type of the preset road; determining a target risk level corresponding to the target risk type based on the second road condition data; based on the target risk type and the corresponding target risk level, determining a target driving strategy when driving on the preset road; and performing driving control on the vehicle based on the target driving strategy. Due to the fact that the first road condition data of the platform data source and the second road condition data of the vehicle-mounted data source are integrated, the road condition of an emergency can be obtained in advance, the corresponding driving strategy is determined, prospective vehicle control is achieved, and the response speed during vehicle control can be increased.

The invention discloses a water conservancy project full life cycle operation and maintenance management method and platform, and relates to the technical field of water conservancy project management.The weir crest water level Hw, the gate flow Qo, the gate opening and closing pressure Pg, the structure dip angle Gq and the iron concentration cFe are accurately collected through a multi-source physical sensing unit; and the sensor type, the arrangement position and the acquisition condition corresponding to each parameter highly accord with the operation characteristics of the water conservancy project, so that the data source is stable, the coverage is wide, and a high-frequency real-time data sensing system is formed. The mechanism breaks through the problem of low credibility of previous parameters, and lays a high-quality data foundation for subsequent operation state modeling and situation map construction. A structural stability correlation function constructed based on a water conservancy physical mechanism is introduced as an evaluation index, sliding window recognition and inversion repair are performed on abnormal fluctuation in original data, and different from a conventional statistical elimination method or an interpolation filling method, reasonable traceability correction can be performed by using a known physical relationship of a system.

The invention relates to the technical field of map data formatting, in particular to a multi-source data fused geological information intelligent surveying and mapping system, which comprises a geological element objectification packaging module, a geological element mapping module, a geological element mapping module, a geological element mapping module, a geological element mapping module, a geological element mapping module, a geological element mapping module, a geological element mapping module, a geological element mapping module and a geological element mapping module, and synchronously obtaining the ID, the timestamp and the data source metadata of each element, and generating a unified geological object set with an identifier. According to the method, objectification integration is carried out on multi-source geological data, and interoperability and traceability are enhanced; a topological relation is internally borne, so that high efficiency of spatial operation and data retrieval is realized; through fine control and encryption of attribute fields and combination of packaging and watermarking technologies, a tamper-proof and traceable closed-loop security guarantee system is constructed, the credibility and integrity of data are improved, and a comprehensive guarantee from generation to verification is formed.

The invention relates to a port micro-grid energy control strategy considering incomplete information, and the strategy is technically characterized in that the time-of-use electricity price of a power grid, the parameters of a micro-grid operator and the like are set, and a data source is determined; scene sampling is carried out, and uncertain information in micro-grid operation is converted into a computable deterministic scene; establishing a Bayesian-Stackelberg game model under incomplete information by taking a power transaction agent as a main body and a micro-grid operator as a slave body; a follower model constraint is established by considering port berth distribution, and a KKT condition is established for a follower model so that the double-layer game model can be converted into a single-layer optimization model; and solving the model by using a CPLEX solver to obtain an optimal electricity price and an energy control strategy of each scene. According to the method, the uncertainty of port micro-grid information is considered, port operation constraints, berth distribution strategies and the like are comprehensively considered, the economy of micro-grid operation is considered while the port logistics transportation efficiency is ensured, and a port micro-grid energy control strategy considering incomplete information is provided.

The invention relates to the technical field of electric power engineering construction, and particularly provides a data processing method and system for scene decision optimization of electric power stringing construction, and the method comprises the steps: constructing a parameter matrix of multiple scenes, and generating a decision task of each scene; for the decision-making task, construction result data is calculated according to a set logic calculation chain through a data flow and logic association engine; the construction result data is displayed through a visual interface, and a working condition input window is provided; updating the parameter matrix, and calculating updated construction result data; and comparing the updated construction result data with a preset safety rule base, performing safety verification, and optimizing the construction scheme based on the parameter matrix under the condition that the verification is not passed. The technical problems of low calculation efficiency, poor accuracy and insufficient safety due to the fact that independent calculation needs to be manually performed according to different data sources in different sub-processes in electric power stringing construction in related technologies are solved.

The invention belongs to the technical field of data watermarks, and particularly relates to a zero-width character-based data watermark embedding and tracing method, which comprises the following steps of: reading a system identifier, to-be-added watermark data and an identity identification number of a corresponding access user; generating watermark information according to the access user identity identification number, the data access timestamp and the system identifier; adopting zero-width characters to encode the watermark information to obtain an encoded zero-width character sequence; splitting the coded zero-width character sequence into a plurality of subsequences based on the identity identification number of the user to obtain a complete subsequence set; and analyzing the to-be-added watermark data, determining to select a corresponding number of embedding positions in the to-be-added watermark data, and embedding the subsequences into the to-be-added watermark data. According to the method, the watermark information and the user identity are strongly bound, unique binding of the watermark, a specific user and an access scene is realized, and a data source can be accurately positioned during data leakage traceability.

The invention discloses an SCRM customer intelligent portraying system and method based on multi-source data, and the system comprises a multi-source data collection and integration module which is used for collecting and cleaning the internal and external multi-source data of an enterprise; the client dynamic label engine is used for generating labels based on business rules and calculating dynamic weights based on data sources and recency; the client intelligent portrait generation and updating module generates a unified portrait through a multi-dimensional portrait model and a portrait fusion device and tracks version evolution; the interactive portrait application and feedback module provides visual display, intelligent early warning and strategy reaching functions and optimizes a label rule through a feedback closed loop; according to the method, internal and external multi-source data of an enterprise are integrated and uniformly coded, so that a data island phenomenon is broken, a multi-dimensional portrait model covering basic identities, transaction values, social behaviors and dynamic intentions is constructed, and a uniform customer panoramic view is formed.

The invention relates to the technical field of brand promotion management, and discloses a promotion management method based on brand evaluation data analysis. The method comprises the following steps: collecting a multi-dimensional evaluation data set of a target brand in a preset period, wherein the multi-dimensional evaluation data set comprises user score data, text evaluation data and market behavior feedback data; performing noise filtering processing on the data set, and extracting an effective evaluation data set with a timestamp and a data source identifier; classifying the data source identifiers, and generating a user side evaluation data subset and a market side feedback data subset; a dynamic weight distribution model is constructed, real-time weight coefficients are distributed for the two subsets based on timestamps, and brand comprehensive evaluation feature vectors are generated through fusion; performing matching retrieval in a historical promotion strategy library according to the vector, and outputting a candidate promotion strategy set; and performing node decomposition on the candidate strategy, extracting an execution condition and an expected influence parameter, generating a verification rule, and screening out a target promotion strategy.

The invention discloses a method and system for constructing a universal digital twin asset data model for multi-source heterogeneous data exchange and sharing, and belongs to the technical field of digital twin. The method comprises the following steps: obtaining three-dimensional data structures and three-dimensional data indexes of different geometric forms in a digital twin model, the spatial topological relation between the three-dimensional data structures is identified; converting texture map information of the three-dimensional data structure into a unified format, normalizing coordinates of the texture map information, and unifying vertex coloring information of the three-dimensional data structure to a standard color space; and performing material analysis and description on different material characteristics of the digital twin model, dynamically calling node objects corresponding to the different material characteristics to obtain real material information, and finally dynamically outputting to obtain a general three-dimensional data model. According to the invention, the problem that a unified digital twin floor scene is difficult to generate due to different sources and structures of current model data can be solved.

The invention relates to a data acquisition unit and method based on an industrial internet unique identifier, and belongs to the technical field of industrial internet and data acquisition, the data acquisition unit comprises a main control module, a communication module and a data acquisition interface module, after equipment is powered on, the main control module firstly automatically initiates a unique identifier registration request to an identifier registration center; the successfully registered identification code is stored in a nonvolatile memory; then, based on the unique identification code, an authentication connection with an industrial internet platform is established through a communication module, and an acquisition configuration instruction issued by the industrial internet platform is received; and generating an acquisition rule according to the acquisition configuration instruction, performing data acquisition through the data acquisition interface module, binding the acquired data with the unique identification code, and reporting the bound data to the platform. According to the invention, global identifiability and credible authentication of the device identity are realized, and dependence on an edge gateway is reduced; the method realizes traceable data source and traceable destination, effectively gets through data islands, and supports cross-domain collaborative application.

The invention relates to the field of mines, in particular to a mine cemented filling body strength intelligent monitoring method. The method comprises the following steps: firstly, preprocessing parameters in obtained data from different data sources to obtain preprocessed parameters; in each observation age, splicing the preprocessing parameters to obtain a fusion feature vector, and obtaining a standardized input sequence; based on the standardized input sequence, a final output feature vector is obtained through an adaptive physical gating multi-scale causal expansion convolution feature extraction algorithm; then, based on the final output feature vector, introducing a double-branch mixed attention mechanism to obtain an attention enhancement feature vector; obtaining a final compressive strength prediction value based on the attention enhancement feature vector; and finally, based on the final compressive strength prediction value, setting a safety early warning mechanism to obtain a monitoring result. The technical problems that the multi-scale receptive field cannot be adaptively adjusted, key stage features are insufficiently captured or redundant calculation is performed, and generalization is poor are solved.